Canceling Closing Nip2 (win-xp)

[MvGulik]

Just making a note of something that I have run into a view times.

On some worksheets, that normally work ok, some error case/state is triggered after "canceling" a "closing nip2" action. (due to my setup I sometimes hit the wrong keys -> [CTRL+Q] instead of [WIN+Q].)

The error remains when trying: -- a general recalculation. -- specific recalculation, or reset, of the error cells. (Could not try much more, as Nips crashed on me while testing something else.)

I'm not sure where or what is triggering this problem. Will try to get some (hopefully more useful) additional information ...

Windows, XP, 32bit, Nip2 7.32.0 (compatibility mode 7.24) -- Not sure if I also have seen this problem on other version. But its probably not limited to the latest version.

[jcupitt]

I've fixed a couple of close-during-recalc bugs in current nip2, perhaps this could be related? There's a test binary here if you have time to check it:

http://www.vips.ecs.soton.ac.uk/development/nip2-7.33.0-setup.piz

rename .piz as .zip, then unzip.

[MvGulik]

Tested this version with two workspaces (A&B) that showed this problem. But compared to 7.32.0 I could not detect any major difference yet. This 7.33.0 test version did however crash in one case(A) on using [CTRL+Q] a second time on the same (not closed) workspace. (compared to no crash on 7.32.0)

Just in case:

• On using the close button on the main Nip2 window these errors don't show up. When I use [CTRL+Q] to close they do show up.
• On one of the two workspaces the errors can be fixed again by "menu: select all + recalculat" (A). (the other needs to be reloaded (B).)

(nts: try to flush out the parts involved here.)

PS: Compatibility information pup-up on loading a older workspace seems to be using the wrong version "This workspace was created by version 7.33.0."

(Windows, XP, 32bit)

[MvGulik]

After taking a better look at the used workspaces, I looks like its a combination of the used compatibility mode and the [Image_file ""] command/function. (I could reproduce it with workspace in compatibility mode with only a single command (a image load).)

File A was using 7.28.1 File B was using 7.24.1

(Time to take a look at the Tabs and other new Nip2 features. :) )

[MvGulik]

Retested with 7.36.0 (Windows, XP, 32bit)

Steps: (can't hurt to do a general rewrite.) -- Opening a (empty/minimal) workspace file that is set to a specific version. -- Adding a general image. (Drag & Drop). -- initiating first [Quit] -> error-state on cell containing the image. (a) -- [Quit] confirmation window: Canceling the quit. -- initiating [Recalculation] -- initiating second [Quit] -> nip2 process crashes. (no windows error message, in case is matters. Nips process just disappears (after, or while, shortly showing a blank message window (the window for confirming the quit action)). (b)

a*) (just data) -- with 7.8 & 7.9 , recalculating will not get rid of the error state. -- with others, recalculating will get rid of the error state.

b*) -- for all compat mode/versions. Hold on ...Yep, the current version (no compat mode) also crashes on this one. (recalculation step not needed, adding image is needed (to, at least, get the confirmation message window).) (Guess that's why this report is still up. ;) )

This report probably should be split up, as a* and b* seem to be different, and probably also unrelated, issues.

[jcupitt]

I can reproduce the first one at least on linux, I'll have a dig. Thanks!

[jcupitt]

I think I've fixed it! Patch here:

https://github.com/jcupitt/nip2/commit/0c8ed95d47190813b6d71d0fe71f08d468d3e7f0

I'll put up a new Windows build soon.

[MvGulik]

Roger.

As I'm lacking the needed compiling skills, I'll have to wait for the Windows build to retest/confirm.

[jcupitt]

OK, there's a 7.36.1 with much better ^Q handling:

http://www.vips.ecs.soton.ac.uk/supported/7.36/win32/

It might fix both problems.

[MvGulik]

Yep, both issues are gone. I can't trigger either of them anymore. :) Thanks.

Closing.